However, since the cognitive revolution, there have been several who have argued for a version of epiphenomenalism. These more recent versions, however, maintain that only the subjective, qualitative aspects of mental states are epiphenomenal. Imagine both Pierre and a robot eating a cupcake. Unlike the robot, Pierre is conscious of eating the cupcake while the behavior is under way. This subjective experience is often called a quale (plural qualia), and it describes the private "raw feel" or the subjective "what-it-is-like" that is the inner accompaniment of many mental states. Thus, while Pierre and the robot are both doing the same thing, only Pierre has the inner conscious experience.
Frank Jackson (1982), for example, once espoused the following view:
I am what is sometimes known as a "qualia freak". I think that there are certain features of bodily sensations especially, but also of certain perceptual experiences, which no amount of purely physical information includes. Tell me everything physical there is to tell about what is going on in a living brain... you won't have told me about the hurtfulness of pains, the itchiness of itches, pangs of jealousy....
According to epiphenomenalism, mental states like Pierre's pleasurable experience—or, at any rate, their distinctive qualia—are just epiphenomena; they are side-effects or by-products of physical processes in the body. Pierre, according to epiphenomenalism, might as well be a robot or a zombie, because conscious mental states do not affect his behavior. If Pierre takes a second bite, it is not caused by his pleasure from the first; If Pierre says, "That was good, so I will take another bite", his speech act is not caused by the preceding pleasure. The conscious experiences that accompany brain processes are causally impotent.
A large body of neurophysiological data seems to support epiphenomenalism. Some of the oldest such data is the Bereitschaftspotential or "readiness potential" in which electrical activity related to voluntary actions can be recorded up to two seconds before the subject is aware of taking a decision to perform the action. More recently Benjamin Libet et al (1979) have shown that it can take 0.5 seconds before a stimulus becomes part of conscious experience even though subjects can respond to the stimulus in reaction time tests within 200 milliseconds. Recent research on the Event Related Potential also shows that conscious experience does not occur until the late phase of the potential (P3 or later) that occurs 300 milliseconds or more after the event. In Bregman's Auditory Continuity Illusion, where a pure tone is followed by broadband noise and the noise is followed by the same pure tone it seems as if the tone occurs throughout the period of noise. This also suggests a delay for processing data before conscious experience occurs. Norretranders has called the delay "The User Illusion" implying that we only have the illusion of conscious control, most actions being controlled automatically by non-conscious parts of the brain with the conscious mind relegated to the role of spectator.
The scientific data seem to support the idea that conscious experience is created by non-conscious processes in the brain (i.e., there is subliminal processing that becomes conscious experience). These results have been interpreted to suggest that people are capable of action before conscious experience of the decision to act occurs. Some argue that this supports epiphenomenalism, since it shows that the feeling of making a decision to act is actually an epiphenomenon; the action happens before the decision, so the decision did not cause the action to occur.
Functionalists chart a different course, accepting that there is a system of mental events mediating stimulus and response, but asserting that this system is "topic neutral" and capable of being realized in various ways. The topic neutrality of the mind implies the denial of epiphenomenalism, which, as a kind of property dualism, fixes consciousness as a non-neutral, non-physical topic.
Eliminative materialists, on the other hand, assert that the concept of mind aims to fix reference to a non-physical topic; so they disagree with the philosophical behaviorist analysis, as well as the functionalist analysis. Eliminative materialism holds, however, that this dualistic aim of "folk psychology" is a fatal error built into mental concepts. So it would be better to eliminate the concept of mind, and concepts implicated in it such as desire and belief, in favor of an emerging neurocomputational account. (A more moderate eliminativist position would maintain what J. L. Mackie called an error theory, stripping false beliefs away from the problematic concepts but not eliminating them, leaving intact a legitimate core of meaning.)
Many argue that data such as the Bereitschaftspotential undermine, rather than support, epiphenomenalism. Such experiments rely on the subject reporting the point in time when conscious experience apparently occurs, which relies on the subject being able to consciously perform an action, and on conscious experience being effective enough to prompt a response. Such a premise contradicts epiphenomenalism, which claims that conscious experience has no effects and therefore cannot be measured. Hence, so the argument goes, any experiment that detects whether or when conscious experience occurs argues strongly against, not for, epiphenomenalism.
Another criticism of epiphenomenalism is that the presence of the theory of epiphenomenalism seems to contradict the very idea. Most would agree that thinking is a mental process, but, if epiphenomenalism is true, how could someone ever express the idea of epiphenomenalism? It would be impossible, because this "expressing" would require the banned connection between mind and behavior. If epiphenomenalism is true and thinking is a mental process, then its truth is ineffable. So in the example above, Pierre cannot convey his pleasure.
Additionally, many argue that the history of epiphenomenalism is revealing. It was concocted as a potential solution to a problem facing dualism: By what mechanism does the mental realm affect the physical? Epiphenomenalism provides an out: The mental realm simply doesn't affect the physical, so the issue is moot. Because it arose out of an attempt to save another conjecture rather than by its own merits, epiphenomenalism can be seen as suspiciously motivated.
Green (2003) has argued that epiphenomenalism does not even provide a satisfactory ‘out’ from the problem of interaction posed by substance dualism. According to Green, epiphenomenalism implies a one-way form of interactionism that is just as hard to conceive of as the two-way form embodied in substance dualism. If it is a problem how mental events can causally influence physical events, how is it any less of a problem how physical events can influence mental ones? Green suggests that the assumption that it is less of a problem may arise from the unexamined belief that physical events have some sort of primacy over mental ones.
If epiphenomenalism is really nothing but a way of rescuing dualism, then the whole issue can be avoided by rejecting dualism. For instance, if the mind is identical to the brain, it must have the same causal powers as the brain, by Leibniz's law.